Vol. 169, No. 10 Role of micF in the tolC-Mediated Regulation of OmpF, a Major Outer Membrane Protein of Escherichia coli K-12 R. MISRAt AND P. R. REEVES* Department of Microbiology, The University of Sydney, Sydney, New South Wales 2006, Australia Received 21 January 1987/Accepted 7 July 1987 Mutation in the toiC locus greatly reduces normal synthesis of OmpF, a major porin protein of Escherichia coli K-l12. 1xperiments that use ompF-ompC chimeric genes demonstrate that a toiC mutation exerts its effect at either the promoter or the amino-terminal end of the ompF gene. Direct analysis of ompF mRNA from tolC+ and toiC strains showed that the amount of ompF transcript in the latter was greatly reduced. We have also observed that, in addition to reducing the amount of OmpF, a toiC mutation increases the levei of OmpC protein to a much greater extent than occurs in an OmpF mutant and also increases micF RNA synthesis as shown by increased j-galactosidase synthesis in a micF-lacZ fusion strain. Based on these observations, we suggest that an increased expression of the mieF gene in a tolC mutant results in the reduced expressioni of omnpF and that a major effect of the toiC mutation may be to push the porin-regulating system to favor ompC and micF to a greater extent than under high-osmolarity conditions. The two proteins, OmpF and OmpC, of Escherichia coli K-12 facilitate diffusion of small hydrophilic molecules through the outer membrane. The amounts of OmpF and OmpC in the membrane vary under different growth condi- tions, with osmotic pressure having a substantial influence: in media of low osmolarity OmpF is synthesized preferen- tially, whereas in high-osmolarity media synthesis of OmpC is preferred (15, 37). The synthesis of these proteins is regulated such that a decrease in the amount of one protein is compensated by an increase in the amount of the other, with the combined amount of the two proteins remaining nearly constant (37). Apart from mutation in the structural genes for OmpF and OmpC, mutations in several other genes can affect the expression of these proteins. Mutation in one such gene, ompB, first characterized by Sarma and Reeves (32), re- sulted in a loss of both OmpF and OmpC proteins. It was later shown that mutation in the ompB locus could result in any of three phenotypes: OmpF- OmpC-, OmpF- OmpC+, or OmpFP OmpC- (32, 38). The ompB locus was further studied by Hall and Silhavy (11, 12), who revealed the presence of at least two genes at this locus: envZ and ompR. OmpR is a cytoplasmic protein and postulated to function as a positive regulatory element (13, 28); the role of EnvZ is less clear. The envZ and ompR genes have been cloned (24) and sequenced (6), but the molecular nature of their role in ompF and ompC gene expression is not clearly understood. Recently a third regulatory locus, micF, was located up- stream from the ompC gene (23). The 3' end of the 173-base- long micF RNA is complementary to the 5' end of ompF mRNA and would be expected to interfere with ompF mRNA translation by forming a stable RNA-RNA hybrid: strains harboring high-copy-number micF+ plasmids in fact lack both OmpF protein and ompF mRNA (23). Mutations in the toiC locus elicit several phenotypic changes, and under certain circumstances toiC mutants lack * Corresponding author. t Present address: Molecular Biology Department, Princeton University, Princeton, NJ 08544. detectable levels of OmpF protein (26). Regulation of this protein by the toiC locus appeared to be independent of that exerted by the ompB locus, because to/C mutants had a similar effect on two other proteins (NmnpC and Lc) (26) which are not under ompB-positive control (30). Experi- ments with an ompF-lacZ operon fusion strain indicated that the tolC-mediated effect on the expression of ompF is at some stage after transcription (26). In this comtnunication we report on studies of the manner in which mutation in the toiC locus affects the expression of the ompF gene. We present data on the strength of the toIC-mediated regulation of OmpF synthesis and compare it with that of the previously characterized ompB regulation. We also studied the effect of toiC on OmpF in micF deletion strains. Our results suggest that the effect of a tolC mutation on OmpF is indirect and involves activation of micF, and this was confirmed by the use of micF-lacZ fusion strains and by use of clones of micE. We have also looked at the interaction of toiC and ompR mutations to further understand the regulation of ompF and ompC expression. MATERIALS AND METHODS Bacterial strains, media, and culture conditions. The bac- terial strains and plasmids used in this study are listed in Table 1. Cultures were grown in our standard nutrient broth medium, being 16 g of nutrient broth (0003; Difco Laborato- ries) plus 5 g of NaCl per liter. This is sometimes referred to as high-osmolarity medium. Low-osmolarity medium was 8 g of nutrient broth (0003; Difco) per liter. When required, the following antibiotics were added: ampicillin (25 ,ug/ml) chlor- amphenicol (25 jig/ml), kanamycin (50 ,ug/ml), and tetracy- cline (16 jig/ml). RNA was labeled in a phosphate-limiting medium which contained 20 mM KCI, 85 mM NaCI, 100 mM Tris, 20 niM NH4CI, 1 mM MgCl2, 0.1 mM KH2PO4, 1 jig of thiamine per ml, 1% Casamino Acids (dephosphorylated), and 5 mg of glucose per ml. Whole-cell envelopes were prepared from cultures grown to late log phase and analyzed by sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS- 4722 JOURNAL OF BACTERIOLOGY, Oct. 1987, p. 4722-4730 0021-9193/87/104722-09$02.00/0 Copyright 3 1987, American Society for Microbiology on October 14, 2020 by guest http://jb.asm.org/ Downloaded from
Transcript
Vol. 169, No. 10
Role of micF in the tolC-Mediated Regulation of OmpF, a MajorOuter Membrane Protein of Escherichia coli K-12
R. MISRAt AND P. R. REEVES*
Department of Microbiology, The University of Sydney, Sydney, New South Wales 2006, Australia
Received 21 January 1987/Accepted 7 July 1987
Mutation in the toiC locus greatly reduces normal synthesis of OmpF, a major porin protein of Escherichiacoli K-l12. 1xperiments that use ompF-ompC chimeric genes demonstrate that a toiC mutation exerts its effectat either the promoter or the amino-terminal end of the ompF gene. Direct analysis of ompF mRNA from tolC+and toiC strains showed that the amount of ompF transcript in the latter was greatly reduced. We have alsoobserved that, in addition to reducing the amount of OmpF, a toiC mutation increases the levei of OmpCprotein to a much greater extent than occurs in an OmpF mutant and also increases micF RNA synthesis asshown by increased j-galactosidase synthesis in a micF-lacZ fusion strain. Based on these observations, wesuggest that an increased expression of the mieF gene in a tolC mutant results in the reduced expressioni of omnpFand that a major effect of the toiC mutation may be to push the porin-regulating system to favor ompC and micFto a greater extent than under high-osmolarity conditions.
The two proteins, OmpF and OmpC, of Escherichia coliK-12 facilitate diffusion of small hydrophilic moleculesthrough the outer membrane. The amounts of OmpF andOmpC in the membrane vary under different growth condi-tions, with osmotic pressure having a substantial influence:in media of low osmolarity OmpF is synthesized preferen-tially, whereas in high-osmolarity media synthesis of OmpCis preferred (15, 37). The synthesis of these proteins isregulated such that a decrease in the amount of one proteinis compensated by an increase in the amount of the other,with the combined amount of the two proteins remainingnearly constant (37).Apart from mutation in the structural genes for OmpF and
OmpC, mutations in several other genes can affect theexpression of these proteins. Mutation in one such gene,ompB, first characterized by Sarma and Reeves (32), re-sulted in a loss of both OmpF and OmpC proteins. It waslater shown that mutation in the ompB locus could result inany of three phenotypes: OmpF- OmpC-, OmpF- OmpC+,or OmpFP OmpC- (32, 38). The ompB locus was furtherstudied by Hall and Silhavy (11, 12), who revealed thepresence of at least two genes at this locus: envZ and ompR.OmpR is a cytoplasmic protein and postulated to function asa positive regulatory element (13, 28); the role of EnvZ isless clear. The envZ and ompR genes have been cloned (24)and sequenced (6), but the molecular nature of their role inompF and ompC gene expression is not clearly understood.Recently a third regulatory locus, micF, was located up-stream from the ompC gene (23). The 3' end of the 173-base-long micF RNA is complementary to the 5' end of ompFmRNA and would be expected to interfere with ompFmRNA translation by forming a stable RNA-RNA hybrid:strains harboring high-copy-number micF+ plasmids in factlack both OmpF protein and ompF mRNA (23).
Mutations in the toiC locus elicit several phenotypicchanges, and under certain circumstances toiC mutants lack
detectable levels of OmpF protein (26). Regulation of thisprotein by the toiC locus appeared to be independent of thatexerted by the ompB locus, because to/C mutants had asimilar effect on two other proteins (NmnpC and Lc) (26)which are not under ompB-positive control (30). Experi-ments with an ompF-lacZ operon fusion strain indicated thatthe tolC-mediated effect on the expression of ompF is atsome stage after transcription (26).
In this comtnunication we report on studies of the mannerin which mutation in the toiC locus affects the expression ofthe ompF gene. We present data on the strength of thetoIC-mediated regulation of OmpF synthesis and compare itwith that of the previously characterized ompB regulation.We also studied the effect of toiC on OmpF in micF deletionstrains.Our results suggest that the effect of a tolC mutation on
OmpF is indirect and involves activation of micF, and thiswas confirmed by the use of micF-lacZ fusion strains and byuse of clones of micE. We have also looked at the interactionof toiC and ompR mutations to further understand theregulation of ompF and ompC expression.
MATERIALS AND METHODSBacterial strains, media, and culture conditions. The bac-
terial strains and plasmids used in this study are listed inTable 1. Cultures were grown in our standard nutrient brothmedium, being 16 g of nutrient broth (0003; Difco Laborato-ries) plus 5 g of NaCl per liter. This is sometimes referred toas high-osmolarity medium. Low-osmolarity medium was 8g of nutrient broth (0003; Difco) per liter. When required, thefollowing antibiotics were added: ampicillin (25 ,ug/ml) chlor-amphenicol (25 jig/ml), kanamycin (50 ,ug/ml), and tetracy-cline (16 jig/ml).RNA was labeled in a phosphate-limiting medium which
contained 20 mM KCI, 85 mM NaCI, 100 mM Tris, 20 niMNH4CI, 1 mM MgCl2, 0.1 mM KH2PO4, 1 jig of thiamine perml, 1% Casamino Acids (dephosphorylated), and 5 mg ofglucose per ml. Whole-cell envelopes were prepared fromcultures grown to late log phase and analyzed by sodiumdodecyl sulfate-polyacrylamide gel electrophoresis (SDS-
4722
JOURNAL OF BACTERIOLOGY, Oct. 1987, p. 4722-47300021-9193/87/104722-09$02.00/0Copyright 3 1987, American Society for Microbiology
pPR268 Apr Tcr; vector, pBR322; cloned This studygene, ompF
pPR272 Kmr; pSC101 replicon; cloned This studygene, ompF
pPR274 Cmr, trpE; mini-F replicon This studypPR275 Cmr; mini-F replicon; cloned This study
gene, ompFpPR426 micF deletion of pMAN006 This studypPR569 4(ompC'-IacZ+)(Hyb) in This study
pPR274
PAGE) as described previously (16, 26). All cultures weregrown at 37°C.DNA techniques. Plasmid DNA was purified by the two-
step CsCl step gradient method of Garger et al. (8). Digestionof plasmid DNA with restriction enzymes, ligation, andtransformation were all performed by standard techniques.DNA fragments were analyzed by electrophoresis in 0.6%agarose gels as described by Maniatis et al. (17). EcoRI-generated fragments of bacteriophage SPPI were used asmolecular weight markers. Nick translation of plasmid DNAfragments extracted from low-melting-point agarose gel wasperformed by the method of Rigby et al. (31).
Digestion of plasmid DNA with restriction enzymes orBal3l, end filling with Klenow fragment, and ligation wereperformed as described by Maniatis et al. (17).A diagrammatic illustration of the subcloning of the ompF
gene into different copy number vector plasmids is shown inFig. 1: pJP33, pPR272, and pPR275 carry ompF in p1SA,pSC101, and mini-F replicions, respectively.
Plasmid pMAN006 contains both micF and ompC genes,and pPR426 (micF- ompC+) was derived from this plasmidby Bal3l. Briefly, pMAN006 was cut at a unique Sall site(located approximately 700 base pairs from the start of themicF gene) and then digested with Bal3l: samples weretaken at various times and the reaction was stopped by theaddition of 5 mM EGTA [ethylene glycol-bis(3-aminoethyl-ether)-N,N,N',N'-tetraacetic acid]. The DNA was incu-bated with Klenow fragment in the presence of all fourdeoxyribonucleotides (dCTP, dATP, dGTP, and TTP) andligated in the presence of phosphorylated BamHI linker(8-mer; pdCGGATCCG). To determine the deletion end-point, a 571-base pair BamHI-EcoRI fragment from pPR426
FIG. 1. Subcloning of the ompF gene into different copy number vector plasmids. The EcoRI-HindIll fragment from pJP33 (ompF inpACYC184 [36]) that carries the ompF gene was inserted between the EcoRI and HindlIl sites of pBR322, resulting in pPR268. TheEcoRI-BamHl fragment from pPR268 was inserted between the EcoRI and BamHl sites of a six-copy-number plasmid, pLG339, resulting inpPR272. To obtain a single-copy ompF plasmid, the chloramphenicol resistance gene of TnI725 (39) was inserted into the HindIll site ofpDF41, resulting in pPR274, and then the EcoRI-SalI fragment from pPR272 was inserted between the EcoRI and Sall sites of pPR274,resulting in pPR275. Abbreviations: B, BamHI; E, EcoRI; H, Hindlll; S, Sall; kb, kilobases.
was subcloned and sequenced from the BamHI linker intoompC.RNA techniques. RNA was purified from exponentially
growing bacterial cultures by the method of Aiba et al. (1)and was further purified by centrifugation in a CsCl gradient(10). Northern transfer of glyoxal-denatured RNA samplesand hybridization with 32P-labeled DNA probe was per-formed essentially as described by Thomas (35). The methodof Parnes et al. (29) was used when a more sensitive assay ofompF mRNA was required. Briefly, about 10 ,ug of purifiedompF DNA (the 513-base pair PstI-PvuII piece from pLF11)was spotted onto a nitrocellulose disk, baked, and hybrid-ized with approximately 5 x 106 cpm of in vivo labeled[32P]RNA (bacteria were grown with [32P]phosphoric acid inphosphate-limiting medium). Hybridized RNA was elutedfrom nitrocellulose filters and, after two phenol extractions,precipitated by ethanol and suspended in 0.1 mM EDTA.The 32P-labeled RNA was then electrophoresed through 5%
acrylamide-8 M urea gels which were autoradiographed at-70°C for 10 to 16 h.
,I-Galactosidase assay. The level of ,-galactosidase infreshly grown cultures was assayed as described by Miller(21).
RESULTS
Analysis of the ompF transcript. To determine the level atwhich a mutation in the toiC locus exerts its effect on ompFexpression, we studied transcription of the omnpF gene bydirectly analyzing ompF mRNA from toiC, ompR101, andompF mutants. RNA was purified from two different parentstrains and their mutant derivatives. These RNA prepara-tions were electrophoresed in an agarose gel and subjectedto Northern transfer to nitrocellulose filters, hybridized witha 32P-labeled DNA fragment of the ompF gene, and autora-diographed. The ompF transcript was present in the parent
ROLE OF inicF IN tolC-MEDlATED REGULATION OF OmpF 4725
CO,
4
0COIna.
Ms0F'Z-O
FIG. 2. Assay of ompF mRNA from wild-type (AB1133), to/C(P602), and ompR101 (P530) strains by the method of Parnes et al.(29). 32P-labeled RNA that hybridized with ompF DNA on filterswas eluted and electrophoresed on a 5% acrylamide-8 M urea geland autoradiographed. As a control, the radioactive RNA isolatedfrom AB1133 was hybridized to a filter with no DNA. Arrowindicates the position of ompF mRNA. A band running above ompFmRNA is the contaminating chromosomal DNA extracted with thecrude labeled RNA preparations.
strains (AB1133 and W1485F-) but was not detected in theto/C, ompF, or ompR101 mutant (data not shown). When amore sensitive RNA-DNA hybridization method (see Mate-rials and Methods) was used, the to/C mutants were shownto have 50-fold less ompF transcript than was present in theparent strain, and none was detected in the ompROlO mutant(Fig. 2).Use of the ompF-ompC chimeric genes to determine the
region of the ompF gene affected by the toiC mutation. Theresults presented above showed that a mutation in the to/Clocus drastically reduces the amount of ompF transcript andpresumably affects the promoter function of the ompF gene.To confirm this, we used chimeric plasmids in which theompF structural gene was placed under ompC promotercontrol or vice versa (18). These and control ompF andompC plasmids were transformed into strains with ompF orto/C mutations either alone or in combination with an ompCmutation. Whole cell envelopes of these strains were pre-pared and analyzed by SDS-PAGE (Fig. 3). Compare P3224and P3226 to see the effect of to/C mutation on ompF underits own ompF promoter control, and compare P3286 andP3231 to see the effect on ompC under the same control.When a gene was under ompC promoter control, the to/Cmutation had no effect (compare P3225 and P3227 for the
effect on ompF and P3285 and P3230 for the effect on ompCexpression). Only when a gene was under ompF promotercontrol was its product reduced by toiC mutation, with thegreater effect being on ompC under ompF promoter control.These results show that the effect of the tolC mutation isexerted at a point upstream of the chimera junction at aminoacid 11 of the mature OmpF protein, on a region whichincludes the promoter and the micF RNA interaction sites.
It should be noted that the toiC effect on OmpF, in strainscarrying chimeric genes, is not as strong as observed in astrain carrying a single copy of the chromosomal ompF gene(e.g., P602). This reduced toiC effect is due to the increase inompF gene dosage, and this aspect is further illustrated in anexperiment described below. At this stage it is not clear whythe toiC effect on OmpC is greater than on OmpF when theseproteins were synthesized under ompF promoter control.
Synthesis of OmpF protein in strains carrying ompF+plasmids of varying copy number. The effect of tolC andompR101 mutations on ompF expression was studied inmutant (toiC or ompR) strains carrying ompF+ plasmidsderived from a mini-F, pSC101, or piSA replicon, which hasan approximate copy number of 1, 6, or 50, respectively.Whole cell envelopes of these strains were prepared andanalyzed by SDS-PAGE (Fig. 4). When OmpF was synthe-sized from the single-copy chromosomal gene, either muta-tion (toiC or ompRI01) reduced the amount ofOmpF proteinbelow the level which could be detected in whole cellenvelopes. However, if the copy number of the ompF genewas increased, the effect of the ompRiOl mutation remainedessentially the same, whereas the toiC mutation was increas-ingly unable to affect the level of OmpF. Thus for 2, 7, and51 copies of ompF, the toiC mutation produced a 20-fold, 4-to 5-fold, and negligible reduction, respectively, in OmpFlevel.
Effect of tolC mutation on OmpF synthesis in micF-ompCand micF deletion mutants. A strain (CS1253) in which theompC and micF genes are deleted was kindly given to us bySchnaitman and McDonald (33). A toiC mutation did nothave the usual dramatic effect on OmpF in this deletionmutant, giving only a two- to threefold reduction in the levelof OmpF when the strains were grown in high-osmolaritymedium and no reduction in strains grown in low-osmolaritymedium (Fig. 5). This experiment indicated that the to/Ceffect on OmpF is largely mediated via the micF or ompCgene or both, but as the absence of the OmpC protein itself
L X U. L 11ILcL > LUU.
XL0 iLOCLLBL CL U.C)L8U.
ct _QLO co Cc LOQ 0La o0'cscm CY C4C % C 00CCOsCOe C-%cz oa c oT-O CM ICYCMcC4eCiNC C4 c'Co
_4 cl es me)3 Coc co coNC4 co co oc
OmPT 4 La. aa a. a.a. a. a.a.Q. a. a. aL a. CL CaOmpC - -__ _ _ _OmpA - _ -_ _
MUTATIONAL ompF tolCBACKGROUND
ompFompC
tolC -tompomp S
FIG. 3. Determination of the region of the ompF gene affected bythe to/C mutation. Strains, with the mutational background indi-cated, were transformed with pMAN007 (FpF), pMAN0O9 (CpF),pMAN006 (CpC), and pMAN010 (FpC), and whole cell envelopesprepared from these strains were analyzed by SDS-PAGE. Only therelevant part of the gel is shown. CpF indicates ompC promoter withompF gene, etc.
does not interfere with the toiC effect on OmpF (see, forexample, P2770 [Fig. 31), it is the micF gene which isimplicated in the suppression of ompF expression observedin toiC mutants.To test this hypothesis, we constructed a Ini(F- ompC4
plasmid, pPR426, from pMAN006 (see Materials and Meth-ods) with a deletion which removed the entire micF gene andended 61 base pairs upstream of the putative -35 region ofthe ompC gene (the BamHI linker was followed by thesequence TACATTTT [23]), leaving the ompC gene intact:an equal amount of the OmpC protein was produced bystrain P3183 (ompF ompC double mutant) carryingpMAN006 or pPR426 (data not shown). Both of theseplasmids were transformed into the micF-ompC deletionstrain CSIZ53 and a toiC derivative, P3398, to give strainseffectively micF+ ompC+ (P3418 and P3419) or micF-ompC' (P3423 and P3424).A comparison of the outer membrane protein profiles of
strains P3418 and P3419, grown in low-osmolarity medium,showed that the tolC mutation in P3419 has the same majoreffect on OmpF level when ompC and mincF are encoded onthe plasmid (pMAN006) as it has when they are on thechromosome. In contrast, comparison of the outer mem-brane protein profiles of strains P3423 and P3424, whichcarry pPR426, showed that in the mnicF- background thetoiC mutation has a negligible effect on OmpF synthesis(data not shown). In high-osmolarity medium the level ofOmpF in P3418 is reduced more than usual relative to
co<Ov-O
o cm
OC) CC0(0LOCDQXC
OnipT N
OmpF-C a
OmpA
OmpTOrmpF-UE----| _
OmpA
OrupTOnipF leOmpC - _ __
OmpF copynumber
One
Two
1 2 34 5 6 7 8
OrnpF*Om
- _ __
OiiipAn
FIG. 5. Effect of toiC on OmpF in micF+ ompC+ and micF-omnpC- strains. Whole cell envelopes from strains W1485F- (lanes1 and 5). P2731 (lanes 2 and 6), CS1253 (lanes 3 and 7), and P3398(lanes 4 and 8) were analyzed by SDS-PAGE. Strains were grown ina low-osmolarity (lanes 1 to 4) or high-osmolarity (lanes 5 to 8)medium.
low-osmolarity medium, presumably due to the additionalcopies of oinpC, but in the micF strain P3423 the reductionis only two- to threefold, with a further four- to fivefoldreduction in P3424, the toiC derivative.
After completion of these experiments, we learned of astrain (SM3001) constructed by Matsuyama and Mizushima(20) in which a deletion specifically removed the entirechromosomal mnicF gene while leaving the ompC gene intact.We are grateful to them for sending us their strain. A toiCmutation in SM3001 caused a four- to fivefold reduction inthe level of OmpF when cultures were grown in a high-osmolarity medium, and in a low-osmolarity medium toiCproduced a two- to threefold reduction in the level of OmpF(Fig. 6). A toiC mutation in the parent micF' strain(MC4100) resulted in almost a total loss of OmpF undersimilar growth conditions (Fig. 6). It should be noted that thedeletion of micEF itself reduces the OmpF level and increasesthe OmpC level, which somewhat obscures the toiC effect onOmpF.
Seven
O nlpF-iWOnipC/
1 _,_b,_ zFiftyOmpAf
FIG. 4. Effect of toIC and oinpR mutations on expression ofOmpF, synthesized by varying copy number plasmids. Whole cellenvelopes from wild-type and mutant strains either without plasmidor harboring ompF+ plasmid pPR275, pPR272, or pJP33 wereanalyzed by SDS-PAGE. Only the revelant part of gels is shown. Itshould be noted that strains carrying multicopy ompF+ plasmidoverproduce OmpF protein and the amount of OmpC protein isconsequently reduced; therefore, it is difficult to resolve the residualOmpC protein from OmpF.
ompF2
OmpFC 23
ompCQrmpA7-
FIG. 6. Effect of toiC on OmpF in the presence or absence of thechromosomal mieF gene. Whole cell envelopes from MC4100 (lane1), P3011 (lane 2), SM3001 (lane 3), and P3493 (lane 4) were analyzedby SDS-PAGE. Strains were grown in a low-osmolarity (A) orhigh-osmolarity (B) medium.
Thus, whereas in low-osmolarity medium tolC had verylittle or no effect on OmpF in micF deletion strains SM3001and CS1253, in a high-osmolarity medium the to/C effect onOmpF was always detectable, but much less than thatobserved in a corresponding micF+ strain.
Effect of tolC on expression of ompC and micF. The resultsof the experiments described above suggest that the to/Ceffect on ompF is in large part mediated by micF. It ispossible that to/C exerts its effect on ompF by varying micFexpression. To study this, a micF-1acZ fusion plasmid(pmicB21) was transformed into wild-type and to/C strains.The expression of micF was assayed by measuring 13-galactosidase activity (Table 2). The to/C mutant had eight-fold more ,B-galactosidase activity than the wild-type strain.The tenfold reduction in micF expression in an ompR mutantconfirms the results obtained by Mizuno et al. (23). Theseobservations clearly show that a mutation in the to/C locusenhances the expression of micF.We find that a to/C mutation, in addition to lowering the
OmpF level, increases the amount of OmpC to a greaterextent than is observed in an ompF mutant, suggesting thatto/C directly affects ompC expression. It has been proposedthat the micF and ompC genes are coregulated (23) and wehave already shown that a to/C mutation increases micFexpression. To provide more direct evidence for the effect ofa to/C mutation on ompC expression, we constructed anompC-lacZ fusion plasmid in which the expression of lacZwas placed under ompC promoter control (Fig. 7). Thisfusion consisted of the ompC promoter, the coding region forthe signal peptide and the first 11 amino acid residues of themature OmpC protein, and the coding region for the LacZprotein frotn residue 8. This fusion was constructed in amulticopy plasmid (pMB1 replicon) to give pPR522 but, dueto the poor or zero growth of strains carrying it, wassubsequently moved into a very low-copy-number plasmid(mini-F replicon) to give pPR569, which was transformedinto wild-type and to/C strains. The expression of ompC wasassayed by measuring ,3-galactosidase activity (Table 2). Theto/C mutant had threefold more ,B-galactosidase activitypresent than the wild-type strain: the ompR mutant, asexpected, had no detectable 1-galactosidase activity. Theseresults agree well with the data to be presented below andconfirm that a to/C mutation increases ompC expression.
Interaction of tolC and ompR mutations. Mutations in theompR gene affect synthesis of one or the other of the OmpFand OmpC proteins. The original ompRiOl mutation resultsin the absence of both OmpF and OmpC proteins; ompR472and ompR20 mutations result in a greatly reduced level ofOmpC, with OmpF synthesized at high levels regardless ofosmolarity in the ompR472 mutant (12), and at a reducedlevel with reversed osmolarity effect in the ompR20 mutant
(27). By comparison, mutation in the to/C locus results in agreatly reduced level of OmpF with constitutive synthesis ofOmpC.
If our oinpC-/acZ fusion plasmid, pPR569, is put into anomnpR472 strain and its to/C derivative, then the ompR472strain has greatly reduced the ,B-galactosidase level as ex-pected, but in the to/C ompR472 strains the level is almost ashigh as in the ompR+ to/C- strain (Table 2): the to/Cmutation appears to have almost overridden the ompR472effect, and the increase due to to/C in this background wasabout 100-fold.The results presented above indicate that to/C interferes
with the normal effect of ompR on OmpC. However, theinterpretation was difficult due to the presence Qf the ompC-lacZ fusion product which interferes with the physiology ofthe cell, and to study the interaction of to/C and ompR undernormal physiological conditions, we used nonfusion strains.The addition of a to/C mutation to two different ompRmutants (MH760, omnpR472; FN101, ompR20) leads to asubstantial increase in the level of OmpC and a decrease inthe OmpF level (Fig. 8). These results confirm out observa-tion that a to/C mutation increases ompC expression.
Effect of tolC in ompF-lacZ fusion strains. We showed
E Sm BI l
FIG. 7. Construction of the oinpC-lacZ fusion. The BglII frag-ment from pMAN006, which includes the ompC promoter and aunique Sall site, was subcloned into the BamHI site of pMC1403,resulting in pPR522. This resulted in a unique fusion of the ompCpromoter to the laccZ gene (at codon 8 of the truncated LacZprotein). The Sall fragment from pPR522 carrying the ompC-lacZfusion was subcloned into the unique Sall site of pPR274, resultingin pPR569. Abbreviations: B, BamHl; Bg, Bgill, E, EcoRl; H.Hindlil; S, Sall, Sm, Smnal.
previously that an ompF-1acZ operon fusion was less sensi-tive to a toiC mutation than the wild-type ompF gene: a toiCmnutation reduces the amount of OmpF below detectablelevels, whereas it reduces ompF expression in the fusiononly by two- to threefold (26). In this study we used an
ompF-lacZ protein fusion strain in which the lacZ gene was
placed under ompF transcription and translation controls(2). The hybrid OmpF-LacZ protein contained the first 35amino acid residues of the amino terminus of OmpF andmost of the LacZ protein. The toiC mutation reduced bothP-galactosidase activity (data not shown) and the amount ofthe hybrid protein by 10-fold in the micF+ strain (Fig. 9,tracks 1 and 2) but not in the mnicF- strain (Fig. 9). TheompR101 mutation reduced ompF expression by >50-fold(data not shown).The protein fusion contains the ompF promoter and the
first 35 amino acid residues of OmpF and hence presumablythe whole of the control region, including the micF bindingsite. In the case of the operon fusion, the fused operon is
known to carry the ompF promoter, but may have none or
only part of the micF binding site. The greater effect of toiCon the protein fusion compared to the operon fusion couldthen be due to differences in response to micF RNA.To test this possibility, we transformed a multicopy micF+
plasmid into ompF-lacZ operon and protein fusion strainsand assayed ompF expression by measuring the 3-
galactosidase activity. The micF plasmid had no effect in theoperon fusion strain (Table 3), whereas it reduced ompFexpression dramatically in the protein fusion strain. Thus,both toiC and high levels of micF produce a significant effectonly in the protein fusion strain, both confirming our hypoth-esis that toliC acts via micF and explaining the low effect oftoiC on the operon fusion studied previously. However, itshould be noted that whereas high levels of micF produce no
effect in the operon fusion strain, a toiC mutation reduces
2 3 4 5 6 7 8
OmpT.OmpC
OmpAC
A...::. ___4_̂'
FIG. 8. Effect of toiC on OmpF and OmpC in ompR and ompR'strains. Whole cell envelopes from MC4100 (lane 1), P3011 (lane 2),MH760 (lane 3), P3394 (lane 4), W4626 Phe- (lane 5), P3396 (lane 6),FN101 (lane 7), and P3393 (lane 8) were analyzed by SDS-PAGE.
1 2 3
.:aa:
FIG. 9. Effect of the toiC mutation on hybrid OmpF-LacZ fusion
and P3289 (MC4100 toiC, lane 3) were analyzed by SDS-PAGE.
Arrow indicates the position of hybrid OmpF-LacZ protein.
ompE expression by two- to threefold (see above). We
believe that this difference reflects a limited toiC effect on
ompE which is independent of micE. A similar limited effect
of toiC on ompE was observed in the micE deletion strains
(Fig. 5 and 6).
DISCUSSIONTolC is a minor outer membrane protein which, in addition
to affecting colicin El tolerance and sensitivity to detergentsand dyes, has a major effect on OmpF, since in a toIC mutantthe amount of OmpF protein present in the outer membraneis dramatically reduced (26). In this paper we show that thiseffect of toiC mutation on OmpF is due to activation of micF,which then inhibits translation of ompF mRNA. The activa-tion of micF may well be mediated by OmpR protein.When we compare the effect of a toiC mutation on strains
which carry different copy numbers of the ompF gene, wefind that in our standard nutrient broth (which includes 0.5%NaCl), a toiC mutation reduces the amount of OmpF proteinin E. coli K-12 to a level not detectable on stained polyacryl-amide gels, whereas it has no effect on OmpF level if thereare 50 copies of the ompF gene present: there was an
TABLE 3. Effect of a multicopy micF plasmid (pCX28) on -galactosidase activity in operon and protein fusion strains
Strain Fusion Plasmid EnzymeUnits
MC4100 0MH513 Operon 418P3625 Operon pBR322 382P3626 Operon pCX28 394MH610 Protein 1,520P3627 Protein pR322 1,426P3628 Protein pCX28 10
ROLE OF micF IN to/C-MEDIATED REGULATION OF OmpF 4729
intermediate reduction if 7 or 2 copies were present. It isclear that ToIC protein is not essential for OmpF synthesis,and this conclusion can also be drawn from our earlierobservation (26) that toIC mutants do produce OmpF proteinin a low-salt medium.Hybrid ompF-ompC genes, constructed by Matsuyama et
al. (18) by in vitro recombination at the common BglII site,enabled us to show that a to/C mutation exerts its effect atthe promoter or at the amino-terminal end of the omnpF gene.This is comparable to the effects exerted via ompR, whichacts on the promoter.We have also observed that, in addition to apparently
reducing the amount of OmpF protein, a to/C mutationincreases the level of OmpC protein present in the mem-brane. This increase is greater than occurs in an ompFmutant (compare, for example, AB1133, P210, and P602 inFig. 3) and hence is not a simple compensation for the lackof OmpF protein, but is presumably an effect of to/Cmutation at the otnpC locus itself. Since we first reportedthat the tolC gene affected expression of the o)mpF gene, ithas been shown (23) that a locus, micF, which maps up-stream of and very close to ompC and is probably coregu-lated with omnpC has a strong negative regulatory effect onompF expression, at least when present on a multicopyplasmid (23). Schnaitman and McDonald (33) also proposeda regulatory element upstream of, and coregulated with,ompC, with a product which inhibited OmpF synthesis.We therefore examined the possibility that the effect of
to/C mutation is primarily on the ompC and micF genes andonly secondarily on the ompF gene.We found that in a micE deletion background a to/C
mutation has virtually no effect on the level of OmpF whencultures are grown in low-osmolarity medium, i.e., nutrientbroth without NaCl: the effect in a high-osmolarity medium(nutrient broth plus NaCI) is difficult to interpret since to/Cmutants are sensitive to high salt levels (unpublished obser-vations), as they are to many other environmental factors,but certainly the effect of to/C is much less in a mi(c mutantthan in a rnicF strain.
Furthermore, to/C mutation is shown to have a differentialeffect on an ompF-/acZ operon fusion and protein fusionswhich correlates with differential sensitivity to a multicopymicF plasmid. The operon fusion may well lack the mni(binding site and is barely affected by to/C mutation, againsupporting the role of mi(E RNA in the to/C effect on OmpF.
It appeared likely, then, that the effect of to/C on OmpF ismediated by activation of the mic-F gene, which has an RNAproduct known to inhibit expression of ompF (23). Thishypothesis is supported by our finding that to/C mutationsubstantially increases the transcription of inicE in a micF-lacZ fusion strain and by our earlier finding (13) that the Stc-mutation, which partially reverses the effect of a to/Cmutation on OmpF protein level (25), is in effect a deletion ofmicF (22). That the major effect of to/C mutation on ompF isat the posttranscription level is quite compatible with it beingmediated by micF RNA, as micF at high copy number hasbeen shown to block translation of ompF mRNA (23).The effect of to/C on ompC was confirmed by use of an
ompC-1acZ fusion. A to/C mutation increases the expressionof ompC, and in particular we show that a to/C mutationincreases omnpC expression when it has been reduced byomnpR mutations. This effect is very clear in the two otmipRmutants used in this study, MH760 (olnpR472) and FN101(ompR20), but was also evident, although not noted at thetime, in a strain (P530) carrying a typical omnpR mutation(ompRI01) which normally lacks OmpF and OmpC (32).
Our data confirm that the expression of minEF and ompC iscoregulated. The results we report support the hypothesisthat the to/C effect on OmpF is brought about by an effect onthis regulatory system, i.e., by affecting expression of theompC and mi(E genes, with the latter then affecting theompF gene.
to/C mutants are pleiotropic and are extremely sensitive todetergents and dyes, indicating that they have a membranedefect: our major results could be explained if this mem-brane defect leads to a modification of the osmolarity detec-tion and response system (which involves ompR and envZgene products) of the cell such that the OmpF/OmpC ratio ispushed even further in favor of OmpC (and ni(cF) than is thecase for a normal strain grown in the high-osmolarity me-dium.
While this would account for the major effect of tolCmnutation on mi(F, ompC, and ompF, we cannot yet accountfor the residual low-level effects on ompF discussed aboveor the previously described effects on NmpC and Lc proteins(26) which do not require OmpR as a positive controlelement (30).Too little is known as yet of the molecular mechanisms
involved in osmoregulation or of the role of omnpR and envZ(9, 19) for us to speculate on the primary effect of ToIC at themolecular level.
LITERATURE CITED
1. Aiba, H., S. Adhya, and B. de Crombrugghe. 1981. Evidence fortwo functional gol promoters in intact Escherichia coli cells. J.Biol. Chem. 256:11905-11910.
2. Berman, M. L., D. E. Jackon, A. Fowler, I. Zabin, L. Christen-sen, N. P. Fiil, and M. N. Hall. 1984. Gene fusion technique:cloning vectors for manipulating lacZ gene fusions. Gene Anal.Technol. 1:43-51.
3. Bolivar, F., F. L. Rodriguez, P. J. Greene, M. B. Betlach, H. L.Heyneker, H. W. Boyer, J. M. Crosa, and S. Falkow. 1977.Construction and characterization of new cloning vehicles. II. Amultipurpose cloning system. Gene 2:95-113.
4. Casadaban, M. J. 1976. Transposition and fusion of the lacgenes to selected promoters in Escherichia coli using bacterio-phage lambda and Mu. J. Mol. Biol. 104:541-555.
5. Casadaban, M. J., A. Martinez-Arias, S. K. Shapira, and J.Chou. 1983. ,B-Galactosidase gene fusions for analyzing geneexpression in Escherichia coli and yeast. Methods Enzymol.100:293-308.
6. Comeau, D. E., K. Ikenaka, K. Tsulng, and M. Inouye. 1985.Primary characterization of the protein products of the Esche-richica coli omnpB locus: structure and regulation of synthesis ofthe OmpR and EnvZ proteins. J. Bacteriol. 164:578-584.
7. Davies, J. K., and P. Reeves. 1975. Genetics of resistance tocolicins in Esclerichia coli K-12: cross-resistance among coli-cins of group A. J. Bacteriol. 123:102-117.
8. Garger, S. J., 0. M. Griffith, and L. K. Grill. 1983. Rapidpurification of plasmid DNA by a single centrifugation in a twostep cesium chloride-ethidium bromide gradient. Biochem. Bio-phys. Res. Commun. 117:835-842.
9. Garrett, S., and T. J. Silhavy. 1987. Isolation of mutations in the(x operon of Esche/richia coli that suppress the transcriptionaldefect conferred by a mutation in the porin regulatory gene. J.Bacteriol. 169:1379-1385.
10. Glisin, V., R. Crkvenjakov, and C. Byus. 1974. Ribonucleic acidisolated by cesium chloride centrifugation. Biochemistry 13:2633-2637.
11. Hall, M. N., and T. J. Silhavy. 1981. The omnipB locus and theregulation of the major outer membrane pore proteins of Escli-eric/ia coli K-12. J. Mol. Biol. 146:23-43.
12. Hall, M. N., and T. J. Silhavy. 1981. Genetic analysis of theo*mpB locus in Escherichia co/i K-12. J. Mol. Biol. 151:1-15.
13. Jo, Y. L., F. Nara, S. Ichihara, T. Mizuno, and S. Mizhushima.
1986. Purification and characterization of the OmpR protein, apositive regulator involved in osmoregulatory expression of theompF and ompC genes in Escherichia coli. J. Biol. Chem. 261:15252-15256.
14. Kahn, M., R. Kolter, C. Thomas, D. Figurski, R. Meyer, E.Remaut, and D. R. Helinski. 1979. Plasmid cloning vehichesderived from plasmids ColEl, F. R6K, and RK2. MethodsEnzymol. 68:268-280.
15. Kawaji, H., T. Mizuno, and S. Mizushima. 1979. Influence ofmolecular size and osmolarity of sugars and dextrans on thesynthesis of outer membrane proteins 0-8 and 0-9 of Esche-richia coli K-12. J. Bacteriol. 140:843-847.
16. Lugtenberg, B., J. MeiJers, R. Peters, P. van der Hoek, and L.van Alphen. 1975. Electrophoretic resolution of the major outermembrane proteins of Escherichia coli K-12 into four bands.FEBS Lett. 58:254-258.
17. Maniatis, T., E. F. Fritsch, and J. Sambrook. 1982. Molecularcloning: a laboratory manual. Cold Spring Harbor Laboratory,Cold Spring Harbor, N.Y.
18. Matsuyama, S. I., K. Inokuchi, and S. Mizushima. 1984. Pro-moter exchange between ompF and ompC, genes for osmoreg-ulated major outer membrane proteins of Escheric hia coli K-12.J. Bacteriol. 158:1041-1047.
19. Matsuyama, S. I., T. Mizuno, and S. Mizushima. 1986. Interac-tion between two regulatory proteins in osmoregulatory expres-sion ofompF and ompC genes in Escherichia coli: a novel ompRmutation suppresses pleiotropic defects caused by an envZmutation. J. Bacteriol. 168:1309-1314.
20. Matsuyama, S. I., and S. Mizushima. 1985. Construction andcharacterization of a deletion mutant lacking micF, a proposedregulatory gene for OmpF synthesis in Escherichia coli. J.Bacteriol. 162:1196-1202.
21. Miller, J. 1972. Experiments in molecular genetics. Cold SpringHarbor Laboratory, Cold Spring Harbor, N.Y.
22. Misra, R., and P. Reeves. 1985. Molecular characterization ofthe Stc- mutation of Escherichia coli K-12. Gene 40:337-342.
23. Mizuno, T., M.-Y. Chou, and M. Inouye. 1984. A uniquemechanism regulating gene expression: translation inhibition bya complementary RNA transcript (mic RNA). Proc. Natl. Acad.Sci. USA 81:1966-1970.
24. Mizuno, T., E. Wurtzel, and M. Inouye. 1982. Cloning of theregulatory genes (ompR and envZ) for the matrix proteins of theEscherichia coli outer membrane. J. Bacteriol. 150:1462-1466.
25. Morona, R., and P. Reeves. 1982. A new locus, stc, whichaffects the phenotype of toIC mutants of Escherichia coli K-12.Mol. Gen. Genet. 187:335-341.
26. Morona, R., and P. Reeves. 1982. The toIC locus of Escherichiacoli affects the expression of three major outer membraneproteins. J. Bacteriol. 150:1016-1023.
27. Nara, F., K. Inokuchi, S. Matsuyama, and S. Mizushima. 1984.Mutation causing reverse osmoregulation of synthesis of OmpF,a major outer membrane protein of Escherichia coli. J. Bacte-riol. 159:688-692.
28. Norioka, S., G. Ramakrishnan, K. Ikenaka, and M. Inouye.1986. Interaction of a transcriptional activator, OmpR, withreciprocally osmoregulated genes, ompF and ompC, of Esche-richia coli. J. Biol. Chem. 261:17113-17119.
29. Parnes, J. R., B. Velan, A. Felsenfeld, L. Ramanathan, U.Ferrini, E. Apella, and J. G. Sidman. 1981. Mouse 12 microglob-ulin DNA clones: a screening procedure for cDNA correspond-ing to rare mRNAs. Proc. Natl. Acad. Sci. USA 78:2253-2257.
30. Pugsley, A. P., and C. A. Schnaitman. 1978. Identification ofthree genes controlling production of new membranes poreproteins in Escherichia coli K-12. J. Bacteriol. 135:1118-1129.
31. Rigby, P. W. J., M. Dieckmann, C. Rhodes, and P. Berg. 1977.Labelling deoxyribonucleic acid to high specific activity in vitroby nick translation with DNA polymerase I. J. Mol. Biol. 113:237-251.
32. Sarma, V., and P. Reeves. 1977. Genetic locus (ompB) affectinga major outer membrane protein in Escherichia coli K,12. J.Bacteriol. 132:23-27.
33. Schnaitman, C. A., and G. A. McDonald. 1984. Regulation ofouter membrane protein synthesis in Escherichia coli K-12:deletion of ompC affects expression of the OmpF protein. J.Bacteriol. 159:555-563.
34. Stoker, N. G., N. F. Fairweather, and B. G. Spratt. 1982.Versatile low-copy-number plasmid vectors for cloning in Esch-erichia coli. Gene 18:335-341.
35. Thomas, P. S. 1980. Hybridization of denatured RNA and smallDNA fragments transferred to nitrocellulose. Proc. Natl. Acad.Sci. USA 77:5201-5205.
36. Tommassen, J., P. van der Ley, A. van der Ende, H. Bergmans,and B. Lugtenberg. 1982. Cloning of ompF, the structural genefor an outer membrane pore protein of E. coli k-12: physicallocalization and homology with the phoE gene. Mol. Gen.Genet. 185:105-110.
37. van Alphen, W., and B. Lugtenberg. 1977. Influence of osmo-larity of the growth medium on the outer membrane proteinpattern of Escherichia coli. J. Bacteriol. 131:623-630.
38. Verhoef, C., B. Lugtenberg, R. van Boxtel, P. de Graaff, and H.Verbeij. 1979. Genetics and biochemistry of the peptidoglycan-associated proteins b and c of Escherichia coli K-12. Mol. Gen.Genet. 169:137-146.
39. Wiebauer, K., S. Schraml, S. W. Shales, and R. Schmitt. 1981.Tetracycline resistance transposon Tnl721: recA-dependentgene amplification and expression of tetracycline resistance. J.Bacteriol. 147:851-859.
